Don't have memory zooms yet, but we do have caching and default start week.

I spent a long amount of today finding various ways that trying to selectively save a user's zoom wouldn't work.

The basics are there, I managed to get the user zoom saved to the page, so the app knows what the user chose before clicking submit time, it's just getting everything lined up so that the app only defaults to the previously selected zip when everything locationally stays the same... throw in the fact that I never really designed the application for that much state awareness from the get go (because I never anticipated someone wanting to save their zoom level) and it's being more difficult than it should and taking much more time than I feel people thought it would.

But, I took a break from that, and got some caching of the NPDS service values working. Now, the service doesn't have to be hit if a call was already made previously, that should help speed things up.

The other thing I implemented was a "default to current week" feature that was requested.

Tomorrow is the deploy. I might brainstorm some nifty way to enable saved zooms before the deploy tomorrow morning, but I can't make any promises. After the deploy I have blocked out some time with other developers to start finding ways to speed up the Quicksilver experience in IE. We have some ideas, we just need to go ahead and try them.

Cheers!

Little Quicksilver Usability updates.

So, now that there is an adjustable legend, I have added a real legend to better explain what the colors mean, and hidden the adjustable bit.

Otherwise, I spent the rest of the day figuring out a way for the app to save the users selected zoom level. That way if they just changed the dates or the search condition, but were zoomed in on a particularly tiny zip3 or zip5, the app would "remember" the last zoom level rather than forcing the user to zoom in from the default again.

Otherwise, I have been talking with Brian about ways to make the application faster. One is to cache averages and standard deviations... and the other is to have a servlet that returns polygon details and eval the polygon additions in a loop eval style. Hopefully this will allow IE users relief from the "This script is taking a long time" error.